Performance Measures

and in shift-workers (Garbarino et al., 2002a). The present study partly supported the

hypothesis that OSA patients would perform more poorly on a test of psychomotor vigilance and a driving simulator task compared to shift-workers, and that shift-workers would perform more poorly on these tasks relative to control participants (hypothesis 3). However, no

significant differences were found between the groups for the number of lapses on the PVT. This finding differs from those reported by Sforza et al (2004) who revealed that relative to control participants, OSA patients exhibited a greater number of lapses on the PVT. It is noteworthy that in the Sforza study, a greater number of lapses were significantly more evident in patients with greater levels of objectively defined sleepiness. Overall, the OSA

118

patients in the present study did not experience significantly high levels of objective sleepiness, and hence, it may have been difficult for an effect to be observed between the groups.

With respect to shift-workers, Axelsson and colleagues (2004) revealed an increase in the number of lapses throughout the night-shift relative to day and evening-shifts. The shift- workers in the present study were tested 24 hours after the conclusion of their last night shift to allow recovery time. It may be that if shift-workers were tested immediately following a night-shift, differences between the groups may have been evident. However, the present study aimed to assess shift-workers at a time when they were not directly affected by sleep periods of acute sleep restriction in order to reflect the longer term effect of shift-work on sleepiness function and mood.

In addition, Bonnefond et al (2006) reported that older age was significantly related to the number of lapses on the PVT. The shift-workers in the present study had a mean age of 45 years compared to 47 years for control participants and 58 years for OSA patients, and this may have concealed differences between the groups, given that they were relatively young. It is also noteworthy to consider that tasks of a longer duration than the PVT (i.e., 10 minutes) (Dinges & Kribbs, 1991) may have produced group differences. Furthermore, the

maintenance of vigilance is largely dependent upon work schedules that promote activity as opposed to monotony (Sallinen et al., 2004), and this was not accounted for in the present study. It is also possible that the low number of participants in each group may have contributed to the lack of differences observed for PVT lapses in the present study. Among other studies which have reported significant group differences, larger sample sizes have been employed. For example, Sforza et al (2004) employed a sample of 152 patients with OSA and 45 control participants and demonstrated that OSA patients had a greater number of lapses

119

than control subjects. Similarly, Axelsson et al (2004) demonstrated an increase in the number of lapses among 317 shift-workers.

OSA patients performed more poorly on median reaction time on the PVT relative to controls and shift-workers respectively, indicating that OSA patients experienced reduced levels of sustained attention. This finding supports previous research which indicates that vigilance is significantly impaired among OSA patients (Beebe et al., 2003; El-Ad & Lavie, 2005). Surprisingly, shift-workers performed better on median reaction time on the PVT relative to control participants in the present study. This finding may be attributable to the time of day that participants were tested. Although night-shift-work has been suggested to produce greater decrements in performance, the shift-workers in the present study obtained 24 hours of recovery time prior to their testing session which was conducted during daytime hours. Hence, it is likely that this may have concealed performance decrements among shift- workers.

With regard to the slowest 10% of reaction times, OSA patients performed the most poorly relative to control participants and shift-workers. In fact, shift-workers fared best on this measure relative to the other groups. Few studies have documented trends in the slowest 10% of reaction times for the PVT among shift-workers. However, previous research has suggested that the performance variability inherent among shift-workers may be greater in the laboratory or testing environment compared with actual operational shift-work settings (Van Dongen, 2006). It may be that this measure of PVT accuracy was more pronounced among shift-workers due to performance variability.

On the driving simulator task, OSA patients exhibited the greatest speed variation from the 70km per hour speed range and the optimal 60 to 80km per hour speed range compared with control participants and shift-workers respectively. This finding partly supported

120

a greater number of errors on measures of speed relative to controls (Findley et al., 1989). However, control participants performed more poorly on measures of speed variability relative to shift-workers. Despite the paucity of research examining driving simulator performance in shift-workers, simulations have been conducted with subjects driving home from the night-shift, and have generally revealed a greater number of incidents, decreased time to accidents, and variability in lateral position when commuting home from the night shift (Akerstedt et al., 2005). The shift-workers recruited for the present study were required to obtain recovery sleep post conclusion of their last night-shift prior to attending for testing. These results suggest that shift-workers are able to perform at the same level as healthy controls during the day after 24 hours of recovery from their night-shift. Moreover, sleep propensity and the circadian drive for sleep is increased during the early morning (Scott et al., 2007). Since participants‟ were tested during daytime hours, the likelihood of driving

performance decrements among shift-workers was likely to be reduced. Furthermore, shifts of longer than 12-hour duration have been suggested to increase the risk of drowsy driving, rendering deteriorations in performance more likely (Scott et al., 2007). Given that the shift- workers in the present study worked an average of 11 hours per day, this may not have been of sufficient duration to observe effects of driving impairment amongst this population.

Average median position from the centre of the left lane, median reaction time and number of crashes did not differ significantly between the three groups. Although these findings differ considerably from previous research which demonstrates an increase in driving simulator impairment in OSA patients and shift-workers relative to control subjects

(Akerstedt et al., 2005; Findley et al., 1995; George et al., 1996; Philip et al., 2005;

Turkington et al., 2001), other characteristics not exclusive to OSA patients may explain these findings. Turkington and colleagues (2001) demonstrated that participant characteristics such as older age, female sex, previous driving experience, previous accident history, and admitted

121

alcohol consumption may influence driving performance. There were few females in the OSA patient group and shift-work group which may explain the inability of the present study to identify differences between the groups. In addition, prior driving experience was not thoroughly examined in the present study. However, the present study contributed to the literature by incorporating a measure of accident history for the preceding three years.

It is noteworthy that one participant was not able to participate in the driving simulator due to feeling disoriented and experiencing dizziness. With this in mind, it may be possible that prior experience or competence with video games may have led to a lack of differences between the groups. Although participants‟ were familiarised with the driving simulator prior

to the 30 minute drive, this may not have been of adequate duration to overcome learning effects with the task. Task duration may also serve to explain the results, since tasks of shorter duration (i.e., 30 minutes on the driving simulator) may not have been sensitive enough to detect group differences. Furthermore, no significant differences were observed between the groups on measures of subjective sleepiness administered immediately after the driving simulator (i.e., the SDQ and SSQ) which indicated that OSA patients and shift-workers reported comparable levels of sleepiness to controls, and may not have been experiencing significant sleepiness during the driving simulator task.

4.3.4. Mood measures. Since the implications of negative mood are often adverse, understanding the nature and extent of such disturbances among patients with OSA and shift- workers seems essential. The present study hypothesised that patients with OSA would demonstrate greater disturbances in mood relative to shift-workers and that shift-workers would report greater mood disturbances compared to control participants and this was partly supported (hypothesis 4). On a specific measure of depression, the BDI, OSA patients endorsed higher scores relative to shift-workers, and shift-workers reported higher scores compared with controls. This finding suggests that OSA patients experienced greater levels of

122

depressive symptomatology compared with shift-workers, and is in line with previous research which suggests that the co-morbidity of depressive symptoms in OSA patients is a common feature of the syndrome (Bardwell et al., 2003; Ferini-Strambi et al., 2003; Pillar & Lavie, 1998; Vandeputte & de Weerd, 2003). Indeed, even when employing different

measures to assess depressive symptoms, these symptoms continue to be observed among this population. In addition, shift-workers demonstrated greater disturbances in depressive

symptomatology as measured by the BDI relative to control subjects in the present study. This finding supports that of previous research which suggests a high prevalence of depressive symptoms during the shift-work experience (Scott et al., 1997). Indeed, higher rates of depression have been indicated for women as opposed to men (Scott et al., 1997).

Interestingly, although the percentage of males in the OSA patient group and shift-work group was much greater than females, differences between the groups were still detected. These findings suggest a strong gender component in the assessment of depression in OSA patients and shift-workers and add further support for the clinical utility of the BDI as a sensitive measure in the assessment of depressive symptomatology among these populations.

On the STAI, no differences were observed between the groups for both State-Anxiety and Trait-Anxiety. The present study attempted to utilise a measure of anxiety that tapped into transitory periods of anxiety as well as more stable personality features that may predispose individuals to more chronic levels of anxiety. Hence, the lack of differences observed between the groups was surprising. Although not statistically significant, inspection of State and Trait Anxiety scores across the three groups revealed that they were in the same direction to that specified in hypothesis 4. Patients with OSA demonstrated higher scores for both State Anxiety and Trait Anxiety relative to shift-workers and controls. It may be the case that the nature of the inventory which asked participants to rate on a Likert scale how they were feeling „right now‟ may have been too limited a time frame to detect differences. Perhaps a

123

measure of anxiety that employed a more liberal time frame for participants to judge their anxiety levels against may have yielded different findings. In addition, it is likely that the small sample size in the present study may have compromised statistical power, rendering group differences difficult to observe.

Since few studies have focused on investigating anxiety in OSA patients and shift- workers, and the present study revealed no differences between the groups on the STAI, further research is needed to document these trends. However, a study conducted by Kjelsberg et al (2005) may explain the lack of differences observed in the present study. Low

compliance with CPAP therapy was found to be the only variable associated with a high level of anxiety among OSA patients. However, OSA patients had only recently received a

diagnosis of OSA, were untreated, and had not yet commenced therapy with CPAP. It may be that effects could not be observed between the groups because OSA patients were not

receiving CPAP therapy. In addition, the present study was cross-sectional and did not assess changes in anxiety after CPAP had been implemented. Future studies may wish to investigate changes in anxiety levels over time after CPAP implementation. In a separate study conducted by Borak et al (1996), sleep fragmentation was the only contributing factor to anxiety levels. However, given that both OSA patients and shift-workers differentially experience

fragmented sleep, this may have led to a lack of differences between the groups.

On the Profile of Mood States, only the Vigour scale and the Total Mood Disturbance score exhibited group differences. Although the scales of Anxiety, Depression, Confusion, Anger, and Fatigue revealed no differences between the groups, there was a trend towards higher scores in both OSA patients and shift-workers. Mosko et al (1989) revealed that the POMS scales did not distinguish patients with OSA from controls. However, it does not seem surprising that the Anxiety scale of the POMS revealed no group differences, since this was also the case for the more specific measure of anxiety, the STAI. Despite this, the lack of

124

differences between the groups for the Depression scale of the POMS was unexpected, given that differences were found for the BDI. Pillar and Lavie (1998) proposed that the POMS Depression scale was a relatively specific measure of depression relative to those more specific to depression, such as the BDI. In addition, the POMS evaluates more transitory moods and general affects when compared with the BDI, perhaps making it more difficult for group differences to be observed.

Although unexpected, no differences were found between the groups for the Confusion scale of the POMS. It could be argued that as a construct, confusion may be unrelated to sleep loss and circadian rhythm disruption. Unlike the constructs of depression and anxiety which appear to be related to poor sleep, the mood state of confusion may be more difficult to disentangle. Surprisingly, the Anger scale of the POMS also revealed no group differences. It could be the case that shift-workers in the present study represented a group of healthy workers who had self-selected into this type of work, and did not experience significant hostility. Similarly, a significant percentage of OSA patients in the present study were retired, and may have been content with their current situation, rendering feelings of anger less likely. As a result of these factors, differences between the groups may have been difficult to detect.

The POMS Fatigue scale also revealed no group differences. However, this finding seems expected since there were no significant differences found for the groups on subjective sleepiness measures. It may be the case that the Fatigue scale of the POMS is not a sensitive measure of fatigue relative to other measures, since it combines a variety of different mood states. On another level, although the terms fatigue and sleepiness are often used

interchangeably to denote the same construct, they each represent a separate concept (Johns, 1998). Fatigue represents a physiological phenomenon characterised by time-on-task

performance decrements which is relieved when the task is changed, however sleepiness refers to an underlying and persistent need state until which is only satisfied when sleep is

125

initiated (Mathis & Hess, 2009). Arguably, measures of subjective fatigue differ from those of subjective sleepiness, rendering group differences difficult to ascertain on the POMS Fatigue scale.

The POMS Vigour scale revealed differences between the groups, supporting hypothesis 4. Specifically and as expected, OSA participants reported the least amount of vigour relative to shift-workers, who in turn reported less vigour relative to controls. Few studies have been conducted examining vigour among OSA patients and shift-workers, however Bardwell et al (1999) found that patients with OSA experienced increased vigour the more they slept. Despite seeming intuitively obvious, this finding supports the notion that when sleep is disturbed, as is the case in OSA patients and shift-workers, vigour is decreased in these populations.

Lastly, the Total Mood Disturbance score of the POMS revealed significant differences between the groups, partially supporting hypothesis 4. Contrary to expectation, shift-workers demonstrated a greater total mood disturbance score relative to OSA patients and controls respectively. Orton and Gruzelier (1989) provided support for the findings obtained in the present study, suggesting that deleterious effects in mood occurred after the night-shift across all mood scales of the POMS relative to controls. Given that the shift-workers in the present study participated after their last night-shift, this may have rendered mood disturbances more pronounced. In terms of OSA patients, demographic characteristics revealed that they worked fewer hours relative to shift-workers and that a significant proportion of patients were retired at the time of the study. These factors may function to explain why shift-workers exhibited higher mood disturbance scores on the POMS relative to OSA patients.